Minutes SIPPING WG at IETF 63 Minutes edited by Gonzalo Camarillo Based on notes by Flemming Adreasen and Stephen Hayes Meeting chaired by Dean Willis, Rohan Mahy, and Gonzalo Camarillo WEDNESDAY, August 3, 2005, 1030-1230 Topic: Agenda Bash Discussions led by: Chairs There was an off-line last-minute discussion between Jonathan Rosenberg and the authors of the profile datasets. As a consequence, the presentation of the profile datasets was moved to the second SIPPING session in order to be able to present the results of this discussion to the group. The chairs presented the status of a number of drafts: o The torture tests draft has problems with the S/MIME examples. The authors will work with the security area director to fix them. o All the URI-list services drafts have been WGLCed already. Since they depend on the consent framework, they will not be sent to the IESG until consent is ready. o There was a problem with the submission of the RTCP summary draft. The authors will add a set of minor comments and resubmit after the meeting. o There were very few comments during the WGLC on the trait-based authorization draft. The chairs polled the room and only few people had read the draft. Among those who had read it, there was consensus that the draft was ready. The chairs will review the draft and decide whether it is ready to be sent to the IESG. o The SIPPING WG assigned Gonzalo Camarillo to perform an expert review on the PoC settings event package draft. He had already performed it before the meeting. Rohan Mahy had concerns related to congestion of non-INVITE-transaction with the multiple-MESSAGE draft. There was a discussion in the room during which it was noted that this had already been discussed one year ago. The conclusion, based on the consensus in the room, was to go ahead with the draft adding some warning text to it. The chairs discussed a number of individual contributions: o The following draft was considered to be out of scope of SIPPING: draft-sohel-sipping-s-bc-concept-arch-00.txt o There were discussions on the following draft: draft-elwell-sipping-redirection-reason-02.txt. The problem of adding additional reason information to responses has appeared many times in the past. The WG is interested studying the problem. We need a solution for it. The chairs will talk to the AD to decide whether or not SIPPING will tackle this work. o The chairs stated that there is no interest and expertise in the WG to work on the following draft: draft-ma-h323-sip-conf-requirement-00.txt. o There were discussions on the following draft: draft-levin-simple-interdomain-reqs-02.txt. Some people stated that this is an interesting problem, but that the solution proposed in the draft is not the right one. Discussions to continue in the mailing list. The chairs will talk to the AD to see if this type of work could fall inside the scope of the WG. o There was consensus in the group that the following draft is useful: draft-hasebe-sipping-exceptional-procedure-example-01.txt. The authors will continue working on it with some volunteers. The chairs will talk to the AD to see whether SIPPING can tackle this as a WG item or whether it should continue as an individual contribution. One thing that will need to be decided is whether examples are enough in all cases or whether we need normative clarifications to some RFCs. o The author of the following draft brought it to the attention of the WG in order to get feedback on it: draft-niemi-sipping-cal-events-00.txt The chairs noted that they are having a lot of problems updating the charter in the official web page. There was a suggestion to maintain an up to date charter on our supplementary page if the problems persist. There was a request for setting up a web page where WGLCs would be scheduled ahead of time so that people could allocate time for them in advance. Gonzalo Camarillo committed to find a secretary for the WG to maintain the web page. Topic: Session-Independent Policies Discussion led by: Volker Hilt Relevant document: draft-ietf-sipping-session-indep-policy-03.txt Slides presented included in the proceedings The author mentioned that the policy format has been aligned with the policy datasets framework. The author asked people to read the draft and send comments. Topic: Session-specific Policy Use Cases Discussion led by: Volker Hilt Relevant document: draft-hilt-sipping-policy-usecases-00-txt Slides presented included in the proceedings The author presented the draft and asked for comments. Topic: Session-specific Policies Discussion led by: Volker Hilt Relevant document: draft-hilt-sipping-session-spec-policy-03-txt Slides presented included in the proceedings The author presented the draft. It was commented that the mechanism needs to take into consideration scenarios with more than one policy server. There were discussions on the mechanism to transport policy information between the user agent and the policy server. It was concluded that more discussions on the list are needed. Topic: Consent Framework Discussion led by: Gonzalo Camarillo Relevant document: draft-ietf-sipping-consent-framework-02.txt Slides presented included in the proceedings The author presented the new draft, which introduces major architectural changes when compared with the previous version of the draft. The author answered questions and clarified some points that were unclear from his presentation. There was consensus in the room that this is the way to go. Topic: SIP-unfriendly Functions in Current Communications Architectures Discussion led by: Jani Hautakorpi Relevant document: draft-camarillo-sipping-sbs-funcs-01.txt Slides presented included in the proceedings The author presented the draft. The common reaction from people that read the draft was that the work is useful and that it should include more details on why things break in the different scenarios presented in the draft. The chairs need to talk to the AD to decide whether we want to use this document as input to other work without publishing as an RFC or whether we publish it as an RFC (when it is ready) to be able to point to the document every time we need to explain somebody why a particular way of implementing a function breaks SIP. Topic: SIP Event Throttles Discussion led by: Aki Niemi Relevant document: draft-niemi-sipping-event-throttle-03.txt Slides presented included in the proceedings The author explained that now the draft describes explicitly how throttles work with RLSs. It can be used without them, but it is meant to be used with RLSs. The author presented a set of open issues. There was consensus that batching functionality should not be included in this document. There was consensus in the room that this is the right direction for this work. The chairs asked whether or not this is ready to be sent to SIP, but there was no clear consensus on this point. Topic: SIP Event Framework Issues Discussion led by: Aki Niemi Relevant document: draft-niemi-sipping-subnot-issues-00.txt Slides presented included in the proceedings There were discussions on the technical issues presented. These discussions will continue on the list. One thing to be clarified in the draft is that the it will deal, in principle, with subscription timeouts, and not with transaction timeouts. Topic: Registration Event Package Extension Discussion led by: Paul Kyzivat Relevant document: draft-kyzivat-sipping-gruu-reg-event-03.txt Slides presented included in the proceedings There was consensus to have the author resubmit the draft as a WG item and WGLC it right after that. The first SIPPING session ended. THURSDAY, August 4, 2005, 1030-1230 Topic: Agenda Bash Discussions led by: Chairs There was no additional agenda bashing besides the one in the first session. Topic: Real-time ToIP Discussion led by: Arnoud van Wijk Relevant document: draft-ietf-sipping-toip-01.txt Slides presented included in the proceedings The author will release a new revision of the draft (02) including a few comments received. There was consensus that revision 02 will be ready for WGLC. Topic: Profile Datasets Discussion led by: Dan Petrie Relevant document: draft-ietf-sipping-profile-datasets-01.txt Slides presented included in the proceedings Jonathan Rosenberg extressed his concerns with the complexity introduced to the model because of data merging. He proposed to focus on data representation and semantics, and tackle merging at a later stage in a different document. A number of people stated that merging should be available from day one. Another issue related to merging that was brought forward was the meta policy needed to implement merging. Coming up with such a meta policy may be an intractable problem in some cases. The discussions will continue on the mailing list. The chairs will talk to Allison to check whether this work is within the scope of SIPPING. There was consensus in the room that, if this work falls within the scope of SIPPING, it should be a WG item. Topic: Proposed Fix to HERFP Discussion led by: Rohan Mahy Relevant document: draft-mahy-sipping-herfp-fix-00.txt Slides presented included in the proceedings There was consensus in the room that this is an important problem the WG should be working on. Regarding the particular mechanism so resolve it, a number of people claimed that there may be alternatives to the mechanism proposed in this draft. Discussions to follow in the mailing list. Topic: TISPAN Discussion led by: Denis Alexeitsev Relevant document: draft-jesske-sipping-tispan-requirements-01.txt draft-jesske-sipping-tispan-analysis-00.txt draft-jesske-sipping-etsi-ngn-reason-00.txt Slides presented included in the proceedings There were discussions on how to proceed with these requirements. The conclusion was to form a design team to clarify the requirements so that the community can understand them better. There was consensus in the room to produce such a clarified requirements document. The design team will not work on solutions, though. Once the requirements are documented the working group will decide the best way to work on solutions for those requirements. Note that previous to this session, there was a TISPAN ad-hoc meeting organized by the SIPPING working group where some of the requirements were discussed. Topic: P-Answer-State Discussion led by: Andrew Allen Relevant document: draft-allen-sipping-poc-p-answer-state-header-00.txt Slides presented included in the proceedings There is an issue on how to indicate in a NOTIFY (related to a REFER) that the answer state is confirmed when the controlling PoC server is at the same time the PoC server of the user agent client. An option seemed to be to place a "fake" P-Answer-State P-header in the body of the NOTIFY. More discussions are needed. Folks will provide Andrew with comments on the mailing list. Once this issue is resolved, SIPPING will perform an expert review on this draft. Topic: Implicit Subscriptions Discussion led by: Jonathan Rosenberg Relevant document: draft-rosenberg-sipping-reg-sub-00.txt Slides presented included in the proceedings There was consensus in the room that there is a problem with implicit subscriptions that needs to be solved. However, people did not agree on where the problem is. More discussions needed in order to figure out which problems are important and which ones are perceived to be important but they are not. Topic: Media Privacy Discussion led by: Henning Schulzrinne Relevant document: draft-shacham-sip-media-privacy-00.txt Slides presented included in the proceedings It was stated that it is hard to determine in an automatic way whether or not a conversation is private. There are two issues: request and indication. Request seems to be solvable. More discussions needed. Topic: SIP User Identifiers Discussion led by: Henning Schulzrinne Relevant document: draft-schulzrinne-sipping-id-relationships-00.txt Slides presented included in the proceedings It was stated that presence can give you the mappings of IDs in a better way. There is a problem when one does not know whether or not the other domain follow these conventions. It was argued that this solution may be usable in enterprise environments. This work was considered to be outside the scope of SIPPING. The second SIPPING session ended.